How to measure employee advocacy without measuring the tool

The dashboard says participation is sixty one percent and the programme is healthy. Then somebody scrolls LinkedIn for two minutes. Three people are posting in their own words about things that were never in the library, and twenty more shared a card in March and have not been seen since. Both facts are true. The dashboard can only see one of them, and it is the less interesting one.

This is how to measure employee advocacy in a way that survives somebody checking. It starts with the distinction that makes most advocacy reporting misleading, and it needs no platform to get going.

TL;DR

Step one: separate the input from the outcome

An advocacy platform works from a content library. Communications puts approved posts in, employees pick one, and the platform records the share. Every figure on the dashboard descends from that single event.

So participation means participation in the library. A colleague who writes their own post every Tuesday contributes nothing to it. Depending on what your programme is for, that is either irrelevant or the whole point, and most programmes have never said which.

Advocacy software counts sharing. Most programmes are trying to encourage writing, which is a different verb and a different number.

The category comparison, sorted by what each platform can observe, is in top employee advocacy software.

Step two: fix the denominator before anything else

Participation rate is the most quoted advocacy number and the easiest to inflate without meaning to, because the denominator is usually the people who enrolled rather than the people who were asked.

Pick the fourth row. A named group who agreed to a number is the only denominator where the resulting percentage means something a manager can act on, and it is small enough that the individual numbers are visible.

Step three: measure the outcome from outside

The outcome measure is simple and it needs nothing installed: what did each person actually publish, on their public profile, this week. It counts a library share and a post written on a phone at a conference identically, because from the outside they are the same object.

1. Posts published per person, per week. Against what they agreed. The whole programme in one column.
2. Written personally against shared from the library. Count these separately or the second will swamp the first.
3. Public likes and comments. Available on any public post, on any account.
4. Who is commenting. Colleagues, peers, or people at target companies. Fifteen minutes, monthly, and worth more than the rest.

The first three can be recorded from the profiles themselves, so somebody who never opted into anything is measured exactly like somebody who did. The argument for running the programme that way is in employee advocacy without an app.

Step four: be careful with reach and earned media value

Advocacy dashboards commonly report a total reach figure and sometimes an earned media value. Both are usually shares multiplied by a follower count and a factor, which is a model rather than a measurement.

Ask for the factor. A vendor with a defensible method will give it in a sentence. Then label the figure in every document it appears in, because the first time somebody senior asks how a number was built is not the moment to find out you cannot say.

Impressions on a personal LinkedIn post are shown to the member who wrote it and to nobody else, including their employer. Any aggregate built from them was either self reported or estimated.

Step five: expect the month four drop, and read it correctly

The first month of a library programme looks excellent. Twenty people share the same card and the totals are enormous. By month four the numbers are down and the standard diagnosis is that adoption fell.

The likelier explanation is that a feed learns. Identical text posted by twenty colleagues reads as a campaign, to the audience and to the ranking alike, so the second and third copies get less attention than the first. A tool that counts shares cannot see that, because it measures the input and the decay happened to the outcome.

This is the strongest practical argument for measuring the outcome somewhere other than the tool that produced the input. Input and outcome measured by the same product will always agree, and agreement is not the same as truth.

Step six: the four lines to report

Whatever you buy or do not buy, this is a reporting pack that survives a busy quarter and a sceptical reader.

• People publishing, against how many agreed to. The programme in one number.
• Posts written personally, counted apart from library shares. The quality signal.
• Comments from named accounts at target companies. Manual, monthly, five minutes.
• What we are not measuring. Reach on personal profiles, and why. Saying it protects every other line.

What has to be true before any number means anything

A measurement of a programme that was never properly set up will be accurate and useless. Three things decide whether there is anything worth counting, and none of them is a tool.

1. A reason for the individual. Nobody posts to raise company reach. They post because it helps their own standing, and the programme has to say so out loud.
2. A number small enough to hold. One post a week beats three, because three collapses in month two and takes the programme with it.
3. Someone senior visibly doing it. A programme where leadership does not publish is a request rather than a norm.

Fix those and a spreadsheet measures the programme adequately. Skip them and the most expensive platform on the market will report, precisely and monthly, that participation is falling. The full version is in how to get employees to post on LinkedIn.

Tell people you are counting

Reading public profiles needs no technical permission, which is exactly why the social permission matters more. A team that finds out it is being counted, rather than being told, will read the whole programme as surveillance and the numbers will be the least of the damage.

Say it plainly at the start: we count posts published, we read nothing private, and the number exists so nobody has to be chased on a Friday. Then write down what is not tracked, which is private analytics, messages and drafts. A short list of what a method cannot see does more for adoption than any dashboard.

Framed as removing the chase it is usually welcomed, because the alternative most people have experienced is a manager asking in person. That framing is also true, which is what makes it hold up when somebody pushes back on it.

A final word on what to do when the numbers are bad. A falling participation rate is usually read as an adoption problem and treated with reminders, which almost never works. Read it instead as a sign that the reason, the number or the example at the top was wrong, and fix one of those three. Reminders move a number for a fortnight and change nothing about why it fell.

The measurement is one part of a wider practice. The whole of it, from the reach arithmetic done honestly to the four ways a programme goes quiet, is in employee advocacy.

How Groowth handles this

It measures the outcome and never the input. You paste the public profile links of the people in the programme and set the cadence each agreed, and it reads those profiles once a day, counting what each person published against what they committed to, with streaks.

There is no library, nothing to approve and nothing to install, so somebody who never opts in is measured exactly like somebody who does, because there is nothing to opt into. It shows no reach and no earned media value, since both would have to be modelled. Three accounts free with no card.

Frequently asked questions